基于MaxEnt的小蓬草在中国的潜在适生区预测

Prediction of Potential Suitable Areas of Conyza canadensis Using MaxEnt Model

小蓬草是我国分布最广的入侵植物之一。根据文献搜集和生物信息数据库,获得有效数据共325个,结合ArcGIS与SPSS相关性分析筛选获得8个气象因子,基于最大熵模型(MaxEnt)预测了小蓬草的潜在适生区。结果表明:基于MaxEnt模型预测小蓬草在中国的潜在适生区ROC曲线,AUC平均值为0.971,预测结果极好;通过刀切法分析表明,最热季度的降水量(BIO_18)、温度季节性变化标准差(BIO_4)、年平均气温(BIO_1)、最冷季度的降水量(BIO_19)4个气象因子对小蓬草的分布影响最大;小蓬草在中国的潜在适生区分布广泛,秦岭淮河以南的各个省份以及秦岭淮河以北至辽宁省南部均为小蓬草高适生区范围。随着气候变化,2050年小蓬草潜在适生区面积与当前相比增加了559016.64 km^(2),2070年小蓬草潜在适生区面积与当前相比增加了627440.30 km^(2)。本研究结果实现对小蓬草入侵动态预警,为进一步防范工作提供了一定的理论基础。

Conyza canadensis is one of the most widely distributed invasive plants in China.According to literature collection and biological information database,a total of 325 valid data were obtained,and 8 meteorological factors were obtained by correlation analysis and screening with ArcGIS and SPSS.The potential suitable areas of the C.canadensis was predicted using MaxEnt model.The results showed that the ROC curve of the potential suitable area of the C.canadensis in China was predicted using the MaxEnt model,and the average AUC was 0.971,which showed excellent prediction results.The jackknife method showed that precipitation in the hottest quarter(BIO_18),standard deviation of seasonal temperature change(BIO_4),average annual temperature(BIO_1)and precipitation in the coldest quarter(BIO_19)had the greatest influence on the distribution of C.canadensis The potential habitat areas of C.canadensis is widely distributed in China,and the provinces south of the Qinling Mountains and Huaihe River and north of the Qinling Mountains and Huaihe River to the south of Liaoning Province are all within the range of C.canadensis high suitable areas.With the change of climate,the area of potential suitable area in 2050 will increase by 559016.64 km^(2),and the area of potential suitable area in 2070 will increase by 627440.30 km^(2) compared with the current.The results of this study can realize the dynamic early warning of the invasion of the C.canadensis,and provide a theoretical basis for further prevention work.